The use of a ground-penetrating radar (GPR) to characterize peat … · 2014-03-28 · 12 th...

12th International Conference on Ground Penetrating Radar, June 16-19, 2008, Birmingham, UK

The use of a ground-penetrating radar (GPR) to characterize peat stratigraphy and estimate the carbon pool in a boreal peatland, Eastmain region, James Bay, Québec, Canada.

Pierre-Luc Dallaire and Michelle Garneau Département de Géographie, Université du Québec à Montréal (UQÀM)

Géotop UQÀM-McGill Montréal, Québec, Canada

emails [email protected] , [email protected]

Abstract -We used a ground-penetrating radar (GPR) to

characterize peat stratigraphy and develop a 3-dimentional

model of an ombrotrophic peatland in the James Bay region,

northern Quebec. GPR measurements have been integrated to

multiple surface topography height measurements realized

with a differential global positioning system (DGPS) while

peat thickness data was obtained by manual probing down to

the mineral surface basin. This method was developed in order

to obtain a more accurate estimate of the total amount of peat

and hence, derive the value of carbon accumulated in the

basin. The objectives of the present study are to : (1)

reconstruct the mineral basin topography underneath the peat

sediment, (2) estimate with the highest precision the

accumulated peat volume stored in the peatland and (3) derive

the total amount of carbon (C) sequestrated. GPR

measurements were realized in March 2007 with a Pulse

EKKO 100 (Sensors and Software), using 100 MHz frequency

antennas. Peat thickness was also measured manually with a

Hiller corer down to the bottom and surface altitude was

obtained with a DGPS 5800/5700 by Trimble. Five peat cores

were collected using a Russian sampler for stratigraphical

description and analyses. With the combination of ArcGIS and

EKKO view Deluxe softwares, the dataset is used to generate a

geographic information system (GIS) in order to reconstruct

spatially and temporally the different paleoenvironment that

contributed to the peat accumulation since 7000 yrs BP. 14C-

dates and loss-on-ignition analyses are used to determine the

carbon values in the different peat layers. The 3D model is

then generated to estimate the peat volume, interpolate the

stratigraphic layers through the basin and calculate the

differential amount of carbon accumulated into these

sequences.

Keywords - peatland, carbon pool, ground penetrating radar (GPR), geographic information system (GIS), James Bay.

I. INTRODUCTION

Peatlands play a key role in carbon (C) cycling because they sequester atmospheric carbon dioxide (CO2) through photosynthesis and release methane (CH4) and CO2 as gas (Gorham, 1991; Francez, 2000; Charman, 2002). The anaerobic conditions prevaling in these ecosystem cause a smaller rate of decomposition by microbial activity, than

the rate of biomass production hence, forming an organic C sink (Vitt et al., 2000). This process allow peatlands to sequester half of the total atmospheric carbon (Cubasch et al, 2001, IPCC). By contrast, if the environmental conditions are modified (e.g. lower moisture content and water table depth, or higher temperature) an increase of decomposition will occur, and therefore, release more greenhouse gas to the atmosphere. Peatlands cover 4 millions km2 on the Earth surface, representing 3% of the total land area. The accumulated peat mass is estimated at 5-6000 Gt (Lappalainen, 1996), wich corresponds approximately to 455 Pg of organic C (Gorham, 1991). This amount represents 1/3 of the organic C stocked in soils of the planet (Francez, 2000). This broad definition concerns peatlands from different bioclimatic regions from the Tropics to the Arctic. Anyhow, approximately 90% of these peatlands are concentrated in the boreal and subarctic regions of the Northern Hemisphere, of which 28% is present in Canada (Francez, 2000). According to Payette and Rochefort (2001), 11.5% of the Québec territories are occupied by peatland ecosystems.

It has became essential in the contemporaneous global warming context to be able to determine with the highest degree of precision of empirical data, the total amount of sequestrated carbon (Gorham, 1991; Kettles and Tarnocai, 1999; Francez, 2000; IPCC, 2007). The use of integrated analyses (e.g. loss-on-ignition (LOI), Troels-Smith or radiocarbon (14C) datation) allows to identify the stratigraphic units forming the peatland sediment. Each of the organic stratigraphic layer possesses unique relative C contents that contribute to the total amount of C stored in the peatland. Previous work have been done to estimate thetotal amount of C using average values for peat thickness, C content as well as bulk density (e.g. Gorham, 1991; Vitt et al., 2000; Sheng et al., 2004). Because the northern Québec peatlands are generally situated in distant zones, obtaining data to generate an efficient database needed to optimize the calculations of C is cost expensive and time consuming. The use of a Ground Penetrating Radar (GPR) represents a good alternative to collect


continuous profiles and cost-effective data (Baraniak, 1983; Warner et al., 1990; Hänninen, 1992; Sheng et al., 2004).

The use of GPR in peatland ecosystems has been documented widely for many practical applications. Baraniak (1983) and Hänninen (1992) used succesfully a GPR device to obtain a continuous reflector for the transition from organic to mineral sediment, and also compare the investigative techniques (GPR versus manual probing) with a significant correlation (respectively a R2 of 0.88 and 0.92). This reflector was also confirmed by Warner et al. (1990) in an ombrotrophic peatland in Ontario (Canada) and by several works lead in Caribou bog, Maine (U.S.A.) (Slater and Reeve, 2002; Comas et al. 2004 and Comas and Slater, 2005). With this technique, stratigraphic units, e.g. moisture content, bulk density or charcoal layers, were successfully distinguished by Warner (1990) and Slater and Reeve (2002). Holden et al. (2002) used GPR to find natural soil pipes in an ombrotrophic peatland with comparing digging trenches and data with a R2 of 0.89 was obtained. Using this geophysical prospection method, Carrier et al. (2002), determined the hydrogeologic relationship between an ombrotrophic peatland and the mineral soil aquifer in order to quantify the water availibility. Many experiences with GPR were lead in Caribou bog, Maine,(Slater and Reeve, 2002; Comas et al., 2004; Comas et al., 2005a; 2005b; Comas and Slater, 2005). Results show that electromagnetic (EM) wave scattering zones corresponded to free-phase biogenic gas (Comas et al., 2005b; Comas and Slater, 2005). The relationship between open pool formation and post-glacial esker presence was found using a GPR system by Comas et al. (2005a). Also, vegetation patterns present on the peat surface were understood by the use of GPR. Comas et al. (2004) and Slater and Reeve (2005) used GPR to demonstrate the link between vegetation patterns and subsurface stratigraphic units which allow them to associate changes in the glacio-marine sediment thickness with surface vegetation assemblages (wooded heath to shrub).

Considering the success of GPR applications in peatland ecosystems, it represents an excellent alternative to increase the database used to estimate the total amount of C in Lac le Caron peatland, James Bay, Québec. The aim of this paper is to present the comparison results from peat volume and peat thickness mean measurements with and without the use of a GPR system.

II. STUDY AREA

Lac le Caron peatland is located in the Eastmain river watershed in the boreal James Bay area, of the northwest portion of the Québec province, 720 km north of the city of Montréal. The coordinates are 52° 17' 6'' N and 75° 50' 19''

W (Figure 1) where mean annual temperature ranges from -2.5 to -5 ºC and mean annual precipitation between 700 and 800mm (Payette and Rochefort, 2001).

The Canadian Shield forms in the region a bedrock structure underneath the peatland, situated more precisely in the Superior Province (4 to 2.5 Ga), mostly formed of igneous rock resulting from the Precambrian geologic period (Brosseau, 2007). The bedrock is covered by several meters of glacial drift accumulated during the different phases of the Wisconsinian glaciations (from 80 Ka to 10 Ka) (Shilts, 1982; Dyke et al. 1989). It was determined, by Dyke et al. (2003), that the sector has been released from the Laurentidian ice sheet around 8.2 Ka BP. The gradual melting of the glacier generated the Tyrrell sea which covered the landscape up to 274 m (Brosseau, 2007). The isostatic rebound, combined with the complete withdrawal of the glacier resulted of the gradual sea retreat. During the Holocene period, the vegetation returned and recolonized gradually the entire region. The Lac le Caron peatland is located at an average altitude of 248 m above msl and the mineral basin underlying the peat mass was formed by sands from deltaic origin, deposited during the withdrawal of the sea (Brosseau, 2007). 14C dating was performed on sediments from the deepest part of the basin in order to get

Figure 1. Localisation of the study area.


the age of the peat initiation in the area (7580 to 7440 cal. BP (# ref. Beta – 223743)).

Poorly diversified vegetation composed mainly of open stands of Picea mariana accompanied with ericaceous shrubs (Chamaedaphne calyculata, Rhododendron groenlandicum) and Carex oligosperma with Sphagnum fuscum and S. magellanicum in the lowest strata) and a low pH oscillating between 3.76 and 4.04 (Garneau et al., 2006) characterizes the Lac le Caron peatland as a bog, or an ombrotrophic peatland (Payette and Rochefort, 2001). This type of peatland is only fed by atmospheric precipitation and isolated from regional groundwater system; therefore a minimum of mineral nutrients are available. Extending over 2.44 km2, the Lac le Caron bog is surrounded to the north, south and east by the stream “Ruisseau à l’eau Claire”, and to the west by a Precambrian outcrop of 290 m above msl. The peatland is characterized by several morphological surface patterns. Most of the complex presents a ridge-hollow pattern (73% of the area) wich consists of an alternance of hollows and hummocks. In the center of the bog, several pools cover 9% of the total surface while in the south of the system, a unit of forested peat dominated by Picea mariana represents 13% of the total area.

III. METHODS

A first series of data was collected during summer 2006. Using a Hiller corer, peat thickness was manually measured from the peat bog surface to the mineral interface. To be considered as a significant measure, mineral sediment needed to be present at the bottom of the sediment or rock should be struck. This was proceed on a 200 m grid covering the whole bog (exempting the northern section, where a 100 m grid was realized) and a total of 117 probes was obtained. Altitude data was collected with a differential global positioning system (DGPS) 5800/5700 by Trimble. A total of 297 altitude points was measured along 8 different transects, at a 25 m interval. Five peat cores were collected using a Russian peat sampler for stratigraphic and other proxies analyses. One of the core was taken at the location of thickest peat sediment while four lateral cores were chosen along the transition to the forested mainland.

A second field campaign was performed in March 2007 to collect the GPR data. These measurements were obtained using a Pulse EKKO 100 from Sensors and Software. A special sledge was built up to protect the instrument from extreme cold temperature (A meteorological tower recorded a minus mean of -36ºC during 2007 winter campaign) and to transport the whole GPR equipment (figure 2). Using a reflection survey mode, four profiles were measured following north-south and east-west transects, with a total lenght of 4.9 km. The different setting parameters used

were: a 100 megahertz (MHz) frequency, a time window of 666 nanoseconds (ns), a station spacing and an antenna separation of 1 m, a sampling interval of 800.48 ns while the antenna orientation was a perpendicular- broadside mode. Using a Garmin GPS Map 76 CS, GPR measures were surimposed on some of the transects realized in the 2006 summer. Thus, it was possible to compare the different type and resolution of data collected at the same geographic point. The spatial distribution of the data picked up during 2006 and 2007 field campaigns are shown in figure 3.

Editing of data was performed using EKKO View Deluxe software. We re-picked time-zero values, merged (and flipped) the sections of the delimited transects, applied a dewow time filter and applied a constant gain of 10. These edits were adopted after Meyers (1996). Using equation (1), we estimated a mean electromagnetic wave velocity in peat. Peat thickness was estimated at each 50m on the transects using equation (2) to upgraded the datset used in several interpolations.

t

pv

)5,0(2 22+

= Equation (1)

vtp )2/(= Equation (2)

Where: υ is velocity

p is peat thickness from Hiller ( meters)

t is two-way travel time (nanoseconds)

Using ArcGIS software, several interpolations, calculations and operations were made. An ordinary Kriging method was used to perform the interpolations, as suggested by Skidmore (2002) and Sheng et al. (2004). This method assumes that proximal measurements are more similar than

Figure 2. Equipment settig for the winter field campaign.


those farther away, and represent a statistically optimal interpolation based on spatial autocorrelation characterized by semivariograms. The data altitude surface was interpolated to obtain a raster where each cell has a unique peat altitude surface value. From this raster, the altitude value was extracted exactly where a peat thickness values was available (Hiller or GPR values). Performing the subtraction between altitude surface and peat thickness, an original value was created wich corresponds to the altitude of the mineral/organic boundary. The interpolation of these values have been made and peat thickness for the whole system obtained by subtracting the surface altitude interpolation to the altitude of the interpolated mineral/organic contact.

In order to confirm the accuracy of these calculations, several statistics analyses were performed. Two sets of 31 measurements of peat thickness were collected with a Hiller corer in a 1 m2 area to estimate the variance of the technique. One set was realized in an unforested section of the peatland (87 % of the area) while the other set was done in a forested section (13 %). The weighted mean was determined to obtain the highest degree of precision. The surface area subdivision was based on the morphological

surface patterns, where unforested corresponds to the ridge-hollow, open pools and structured fen sections. To estimate accuracy of peat thickness interpolation accuracy, several data subsets were extracted randomly from the originals databases. For each peat thickness data, five different extractions including 5 % of the original database values, were realized. The sequence of operation leading to the interpolation was done with these different data subsets. The comparison between measured and predicted data was performed using a linear regression correlation. This subsets technique is based on Sheng (2004) and Tapsoba et al. (2005).

IV. RESULTS

1.1 Estimation with Hiller corer data

The statistical analysis reveals a variance of ±0.057 cm awarded to the Hiller corer, and an uncertainty in the average of 0.01 cm. Figure 4 presents the interpolation of peat thickness using the data from the Hiller corer measurements. The comparison of the measured and predicted data for peat thickness calculation using only the Hiller corer (n= 28) was obtained with a R2 of 0.68 (ρ= <0.0001). The mean of the interpolated values for peat thickness is 2.55 m which is closed from the mean of the measured peat thickness calculated with measured values (2.56 m). Using only the manual probe measurements, peat volume was estimated to be 7.8x106 m3 (calculated with 3D analyst in ArcGIS software).

Figure 3. Spatial distribution of the data into the Lac le

Caron peatland.

Figure 4. Interpolation of peat thickness with Hiller corer

measurements.


1.2 Estimation with Hiller corer and GPR data

Using equation (1), 39 control points were utilised to calculate a mean electromagnetic wave velocity in peat at 0.046 ns/m. A R2 of 0.88 (ρ= <0.0001) was obtained with a linear correlation between manual and GPR peat thickness measurements. Figure 5 presents the manual and GPR curves formed with all 39 control points and allows to distinguish the relationship between these peat thickness measurements. Figure 6 shows the interpolation realized with Hiller corer and GPR data. The correlation used to quantify the validation of this interpolation obtained a R2 of 0.77 (n= 60) (ρ= <0.0001). The mean peat thickness obtained with this interpolation is 2.63m while the dataset used had a mean of 2.57m. The calculation that was based on the data from two different techniques, estimated the peat volume of 7.9x106 m3.

1.3 Spatial distribution of the peat thickness

The two interpolations present some differences and similarities on the spatial distribibution of peat thickness (Figures 4 and 6). Both maps present the thickest section at the boundary correspondding to the outcrop (West side). From this point, peat thickness decreases gradually to reach a minimal thickness along the shore of the “Ruisseau à l’eau Claire”. When compared to Figure 4, which presents interpolated data with Hiller corer, Figure 6 (Hiller and GPR data) shows a more detailed representation of peat thickness variations. The largest ones being located at the North-West, South-East section and in the center North of the peatland.

Figure 5. Comparisson between the manual and GPR curves.

Figure 6. Interpolation of peat thickness with Hiller corer

and GPR measurements.

-6

-5

-4

-3

-2

-1

0

1 8 15 22 29 36

Control points

De

pth

(m

)

manual probes

GPR


V. COMPARISON

1.1 Microtopography scale

Figure 7 combines two illustrations demonstrating the higher precision offered by GPR data to identify the peat/mineral boundary. Part A presents the microtopography from the LLC_E-O(3)_100 GPR profile (Figure 2, transect West-East in the southern part of the peatland). Part B shows on the same transect, the interpolated curves between the measured points from both techniques with a 200m interval with cored data and 50m interval for GPR data. Offering a continuous data acquisition, GPR allows a higher precision the microtopography underneath the peat. Using the Hiller corer dataset only, should hide some microtopographic elements which alters the estimation of the peat volume, and impairs the C calculation.

1.2 Peatland scale

Calculating the peat volume with the interpolation using

GPR data, an increase of 115 175m3 was recorded. The mean thickness of the interpolation increases as well from 2.55m to 2.63m. Those differences are also recorded in figure 8 which prensents the subtraction between Hiller data interpolation and Hiller and GPR data interpolation (respectively fig. 4 and 6). The figure confirms the spatial differences between the interpolations where the highest differences are exactely located where GPR transect was realized. The whole peatland complex has then been redefined to precise peat volume estimation and carbon evaluation.

1.3 C calculation and statistical validations

Based on the formula suggested by Sheng et al. (2004) to calulate C contents in a peatland, a difference of 11 792 metric tons appears using the mean thickness method from the Hiller and GPR interpolation (2.63m). This confirms that GPR data allows a more accurate estimation of peat volume, mean peat thickness and related total C

Figure 7. (A) LLC_E-O(3)_100 GPR profile, (B) the interpolated curves between the measured points from both techniques.

-3

-2,5

-2

-1,5

-1

-0,5

0

-50 50 150 250 350 450 550

D i st a nc e a long pr of i l e LLC _ O- E( 3 ) _ 10 0 ( m)

manual

gpr B

A

No

data

obt

aine

d

Peat

thic

knes

s (m

)


Figure 8. Results of the difference between Hiller data

interpolation and Hiller and GPR data interpolation.

sequestrated. With the addition of GPR data, the R2 of the linear correlation between measured and predicted data increased from 0.68 to 0.77. GPR data upgraded the validation of the interpolation and allows to consider data with a more confidence degree. By contrast, the difference between the mean measured point thickness and mean thickness by interpolation is higher with data from Hiller and GPR (6cm versus 1cm) then decreasing the certainty of the model.

VI. CONCLUSION

GPR thickness data are significantly similar to manual measurements and from our results are recommended to be used to increase the dataset accuracy. Interpolation with GPR data has proven to be very effective while offering a more precise mean of peat thickness, peat volume, sequestrated C calculation with a strong validation coefficient. GPR continuous data allows a better microtopographic reconstruction. Considering the time saved by GPR data collecting, it was demonstrated that this geophysical prospective method can be succesfully used to complete manual probing in order to estimate peat thickness. Future work will help in determining also if the

use of GPR will determine higher precision of the bulk density and the moisture content o the peat sediments.

ACKNOWLEDGMENTS

This research could not have been realized without the logistics and financial support provided by Hydro-Québec (Unité Production), especially Dr. Alain Tremblay, senior scientific project leader. We also want to thank people who helped during the field campaigns: Éric Rosa, Julie Loisel, Laurence Parenteau and Simon van Bellen. Special thanks is dedicated to Éric Rosa for his invaluable help during the GPR comprehension.

REFERENCES

Baraniak, David W., and Donohue & Associates. 1983. «Exploration for surface peat deposits using ground penetrating radar». In (C.H. Fuchsman and S.A. Spigarelli eds) Symposium on peat utilization (October 10-13 1983).

Brosseau, Daniel. 2008. «Caractérisation des dépôts meubles et reconstitution paléogéographique Quaternaire de la région du réservoir Eastmain-1, baie James, Québec». Montréal, Mémoire M.Sc., Département de Géographie, Université du Québec à Montréal, 120 p.

Carrier, Cynthia, Yves Michaud, Allan S. Crowe and Michel Allard. 2002. «Caractérisation hydrogéologique d'un tourbière ombrotrophe dans le sud-est du Nouveau-Brunswick: résultats-préliminaires». In IAH-CGS Groundwater Specialty Conferences (Ground and water: Theory to Practice) (October 20-23, 2002): D, Stoll, A.R. Piggott and J.J. Crowder (Southern Ontario Section of the Canadian Geotechnical Society).

Charman, Dan. 2002. Peatlands and environmental change. Chichester: Willey, 301 p.

Comas, Xavier, and Lee D. Slater. 2005. «Geophysical and hydrological evaluation of two bog complexes in a northern peatland: Implications for the distribution of biogenic gases at the basin scale». Global Biogeochemical Cycles. Vol. 19, p. 1-10.

Comas, Xavier, Lee D. Slater and Andrew Reeve. 2004. «Geophysical evidence for peat basin morphology and stratigraphic controls on vegetation observed in a Northern peatland». Journal of hydrology. Vol. 295, p. 173-184.

Comas, Xavier, Lee D. Slater and Andrew Reeve. 2005 a. «Spatial variability in biogenic gas accumulations in peat soil is


revealed by ground penetrating radar (GPR)». Geophysical research letters. Vol. 32, p. 1-4.

Comas, Xavier, Lee D. Slater and Andrew Reeve. 2005 b. «Stratigraphic controls on pool formation in a domed bog inferred from ground penetrating radar (GPR)». Journal of hydrology. Vol. 315, p. 40-51.

Cubasch, U., G. A. Meehl, G. J. Boer, R. J. Stouffer and M. Dix. 2001. «Projections of future climate change ». In Climate Change 2001: The Scientific Basis., J. T. Houghton, Y. Ding, D.J. Griggs, M. Noguer, P. J. van der Linden, X. Dai, K. Maskell and C. A. Johnson (Eds.). Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom, p.526-582.

Dyke, A. S., J.-S. Vincent, J. T. Andrews, L. A. Dredge and W. R. Cowan. 1989. «L'inlandsis laurentidien: Introduction à la géologie quaternaire du Bouclier Canadien». In Le Quaternaire du Canada et du Groenland,Géologie du Canada, R.J. Fulton, p. 188-202. Canada: Commission géologique du Canada.

Dyke, A. S., A. Moore and L. Robertson (2003). Deglaciation of North America. Canada: Commission géologique du Canada Dossier public 1574

Francez, André-Jean 2000. «La dynamique du carbone dans les tourbières à Sphagnum, de la sphaigne à l'effet de serre». Année Biologique. Vol. 39, p. 205-270.

Garneau, Michelle , Hans Asnong, Véronique Beaulieu-Audy, Daniel Brosseau, Robert-André Daigneault and Luc Pelletier. 2006. Rapport des travaux 2005 Caractérisation des milieux terrestres avant la mise en eau du réservoir EM-1. Rapport pour Hydro-Québec. Montréal, Université du Québec à Montréal: 144 p. (www.eastmain1.org).

Gorham, Eville 1991. «Northern peatlands:role in the carbon cycle and probable responses to climatic warning». Ecological Applications. Vol. 2, p. 182-195.

Hänninen, Pauli. 1992. «Application of ground penetrating radar and radio wave moisture probe techniques to peatland investigation». Geological survey of Finland. Bulletin 361, p. 1- 71.

Holden, J., T.P. Burt and M. Vilas. 2002. «Application of ground-penetrating radar to the identification of subsurface in blanket peat». Earth Surface Process and Landforms. Vol. 27, p. 235-249.

IPCC, (intergovernmental Panel on Climate Change). 2007. «Climate Change 2007: AR4 Synthesis Report». An assessement of the Intergovernmental Panel on Climate Change. Watson, R. T. et al. (Eds.). Cambridge University Press. 73 p.

Kettles, Inez M., and Charles Tarnocai. 1999. «Development of a model for estimating the sensitivity of Canadian peatlands to climate warming». Géographie physique et Quaternaire. Vol. 53, no 3, p. 323-338.

Lappalainen, E. 1996. «General review on world peatland and peat resources». In: E. Lappalainen (Eds) Global peat ressources, UNESCO, International Peat Society, Geological Survey of Finland, p. 53- 56.

Meyers, Richard A., Derald G. Smith, Harry M. Jol and Curt D. Peterson. 1996. «Evidence for eight great earthquake-subsidence events detected with ground-penetrating radar, Willapa barrier, Washington». Geology. Vol. 24, no 2, p. 99-102.

Payette, Serge, and Line Rochefort (Eds). 2001. Écologie des tourbières du Québec-Labrador. Saint-Nicolas: Les Presses de l'Université Laval, 644 p.

Sheng, Yongwei, Laurence C. Smith, Glen M. MacDonald, Konstantine V. Kremenetski, Karen E. Frey, Andrei A. Velichko, Mary Lee, David W. Beilman and Peter Dubinin. 2004. «A high-resolution GIS-based inventory of the Siberian peat carbon pool». Global Biogeochemical Cycles. Vol. 18, p. 1-14.

Shilts, W. W. 1982. «Quaternary evolution o the Hudson / James Bay region». Le Naturaliste canadien. Vol. 109, p. 309-332.

Skidmore, Andrew. 2002. Environmental modelling with GIS and remote sensing. New York: Taylor and Francis, 268 p.

Slater, Lee D. , and Andrew Reeve. 2002. «Investigating peatland stratigraphy and hydrogeology using integrated electrical geophysics». Geophysics. Vol. 67, no 2, p. 265-378.

Tapsoba, Dominique, Vincent Fortin, François Anctil and Mario Haché. 2005. «Apport de la technique du Krigeage avec dérive externe pour une cartographie raisonnée de l'équivalent en eau de la neige: Application aux bassins de la rivière Gatineau». Canadian journal of civil engineering. Vol. 32, p. 289-297.


Vitt, Dale H., Linda A. Halsey, Ilka E. Bauer and Celina Campbell. 2000. «Spatial and temporal trends in carbon storage of peatlands of continental western Canada through the Holocene». Canadian Journal of Earth Science. Vol. 37, p. 683-693.

Warner, Barry G. , David C. Nobes and Brian D. Theimer. 1990. «An application of ground penetrating radar to peat stratigraphy of Ellice Swamp, southwestern Ontario». Canadian Journal of Earth Science. Vol. 27, p. 932-938.

The use of a ground-penetrating radar (GPR) to characterize peat … · 2014-03-28 · 12 th...

Documents

Transcript of The use of a ground-penetrating radar (GPR) to characterize peat … · 2014-03-28 · 12 th...